Abstract
Few prospective studies have examined associations between diet quality and pancreatic ductal adenocarcinoma (PDAC), or comprehensively compared diet quality indices. We conducted a prospective analysis of adherence to the Healthy Eating Index (HEI)-2015, alternative HEI-2010, alternate Mediterranean diet (aMed), and 2 versions of Dietary Approaches to Stop Hypertension (DASH; Fung and Mellen) and PDAC within the National Institutes of Health (NIH)-AARP Diet and Health Study (United States, 1995–2011). The dietary quality indices were calculated using responses from a 124-item food frequency questionnaire completed by 535,824 participants (315,780 men and 220,044 women). We used Cox proportional hazards regression models to calculate adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) for each diet quality index and PDAC. During follow-up through 2011 (15.5-year median), 3,137 incident PDAC cases were identified. Compared with those with the lowest adherence quintile, participants with the highest adherence to the HEI-2015 (HR = 0.84, 95% CI: 0.75, 0.94), aMed (HR = 0.82, 95% CI: 0.73, 0.93), DASH-Fung (HR = 0.85, 95% CI: 0.77, 0.95), and DASH-Mellen (HR = 0.86, 95% CI: 0.77, 0.96) had a statistically significant, lower PDAC risk; this was not found for the alternative HEI-2010 (HR = 0.93, 95% CI: 0.83, 1.04). This prospective observational study supports the hypothesis that greater adherence to the HEI-2015, aMed, and DASH dietary recommendations may reduce PDAC.
Keywords: AHEI-2010, aMed, DASH, diet, HEI-2015, pancreatic cancer
Abbreviations
- AHEI-2010
Alternative Healthy Eating Index-2010
- aMed
alternate Mediterranean diet
- CI
confidence interval
- DASH
Dietary Approaches to Stop Hypertension
- FFQ
food frequency questionnaire
- HEI-2015
Healthy Eating Index-2015
- HR
hazard ratio
- NCI
National Cancer Institute
- NIH
National Institutes of Health
- PDAC
pancreatic ductal adenocarcinoma
Although pancreatic cancer is relatively rare and accounts for only 3% of incident cancer cases in the United States, it is among the most lethal of all major cancers, with a 5-year survival rate of only 10% (1). Pancreatic ductal adenocarcinoma (PDAC) is the most common pancreatic cancer type and accounts for more than 85% of pancreatic cancers (2). Potentially modifiable risk factors for PDAC include cigarette smoking, excess body weight, type 2 diabetes mellitus, and diet (3). In studies of individual nutrients or foods and PDAC risk, the most consistently reported associations have been for higher PDAC risk with heavy alcohol use (4–6) and inconsistent associations for higher consumption of red meat and dietary fat (7–10).
In contrast to individual foods and nutrients, dietary patterns can account for complex correlations and interactions that are not detected when evaluating associations for individual foods or nutrients (11). The Dietary Patterns Methods Project identified the 4 most commonly used a priori–defined US diet quality indices: the Healthy Eating Index (HEI) (12, 13), based on the Dietary Guidelines for Americans (14); Alternative HEI (AHEI) (15), based on Harvard’s Healthy Eating Plate (16); alternate Mediterranean diet score (aMed) (17), based on the Mediterranean Diet (18); and Dietary Approaches to Stop Hypertension (DASH) (19), based on the DASH Eating Plan (20–22). These patterns emphasize higher consumption of fruits, vegetables, whole grains, and legumes and limited consumption of refined grains, red and processed meats, sugar-sweetened beverages, added sugars, and saturated fats. Accumulating evidence suggests that greater adherence to these diet quality indices is associated with lower risk of cancer incidence and mortality (23, 24).
Three prospective studies have evaluated the association between aMed and HEI-2005 indices and pancreatic cancer risk with conflicting results (25–27). Since the publication of the earlier studies of diet and PDAC risk within National Institutes of Health (NIH)-AARP (formerly the American Association of Retired Persons) (26, 28), there has been longer follow-up and more incident PDAC cases. To compare variations between diet indices and PDAC risk, we examined the associations between adherence scores to 5 sets of diet quality index recommendations. To be consistent with the Dietary Patterns Methods Project (24), in this analysis, we considered the HEI-2015 (12, 13), AHEI-2010 (15), aMed (17), and 2 DASH diet indices, one based on food groups (Fung et al.) (19) and the other based on nutrients (Mellen et al.) (29). To the best of our knowledge, HEI-2015, AHEI-2010, and the 2 DASH scores have not previously been examined and compared in relation to PDAC risk. We hypothesized that greater adherence to diet quality indices would be associated with lower PDAC risk.
METHODS
Study population
The NIH-AARP Diet and Health Study is a large prospective cohort of male and female AARP members, aged 50–71 years at baseline, who resided in 1 of 6 states (California, Florida, Louisiana, New Jersey, North Carolina, and Pennsylvania) or 2 metropolitan areas (Atlanta, Georgia, and Detroit, Michigan) (30). During 1995 and 1996, self-administered questionnaires queried participants to provide information about dietary intake during the previous 12 months, demographic characteristics, and health-related behaviors, including physical activity and smoking status (30). In total, 566,398 participants satisfactorily completed and returned the questionnaires. The NIH-AARP Diet and Health Study was approved by the Special Studies Institutional Review Board of the US National Cancer Institute (NCI), and all participants gave informed consent.
Participants whose questionnaires were completed by a proxy or who had a prevalent history of cancer (except nonmelanoma skin cancer) based on cancer registry data, end-stage renal disease, or reported extreme energy intake (2 interquartile ranges below the sex-specific 25th percentile or above the 75th percentile of log-transformed energy intake) or with person-years ≤0 were excluded. Our analytical sample included 535,824 participants (315,780 men and 220,044 women; Web Figure 1, available at https://doi.org/10.1093/aje/kwac082).
Dietary assessment and index-based dietary quality indices
Participants completed a self-administered semiquantitative 124-item food frequency questionnaire (FFQ) that queried frequency and portion size of foods and beverages over the previous 12 months (30). Further validation of the FFQ was performed within a subset of the NIH-AARP Diet and Health Study, using 2 24-hour dietary recalls (31). Details regarding the FFQ are published elsewhere (31–33).
The dietary data from the FFQ was linked to the MyPyramid Equivalents Database (MPED), version 1.0, to derive guidance-based food group equivalents for whole grains, total grains, total vegetables (including all vegetable subgroups), total fruit, low-fat dairy, protein foods (including poultry, fish, nuts, soy, and legumes), solid fat, added sugars, and alcohol. Additionally, nutrient estimates were generated for saturated fat, monounsaturated fat (MUFA), polyunsaturated fat (PUFA), eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA), trans-fat, sodium, and alcohol by using the US Department of Agriculture Survey Nutrient Database associated with Continuing Survey for Food Intake by Individuals 1994–1996 and the Nutrition Data System for Research. The MPED and nutrient variables were used to create the dietary quality indices (34) for the HEI-2015, AHEI-2010, aMed, and DASH.
The components for the 5 dietary indices are summarized in Table 1, and their scoring is described in the Web Appendix. Briefly, the HEI-2015 consisted of 13 components with a range of 0–100 points (13), the AHEI consisted of 11 components (range, 0–110 points) (15), the aMed consisted of 9 components (range, 0–9 points) (17), the DASH-Fung consisted of 8 components (range, 0–40 points) (19), and the DASH-Mellen consisted of 9 components (range, 0–9 points) (29).
Table 1.
Standards for Maximum Scores Across 5 Dietary Quality Indices in the National Institutes of Health–AARP Diet and Health Study, United States, 1995–2011
| Diet Quality Indices | |||||
|---|---|---|---|---|---|
| Diet Quality Component | HEI-2015 | AHEI-2010 | aMed a | DASH-Fung b | DASH-Mellen |
| Adequacy components | |||||
| Calcium | ≥590 mg/1,000 kcal | ||||
| Fiber | ≥14.8 g/1,000 kcal | ||||
| Magnesium | ≥238 mg/1,000 kcal | ||||
| Potassium | ≥2,238 mg/1,000 kcal | ||||
| Protein | ≥18% kcal/day | ||||
| MUFA:saturated fat ratio | ≥ median: M: 1.2; W: 1.2 | ||||
| PUFA | ≥10% kcal/day | ||||
| PUFA+MUFA:saturated fat ratio | ≥2.5 | ||||
| EPA+DHA | 250 mg/day | ||||
| Fruits | ≥0.8 cup eq/1,000 kcal | ≥ median (cup eq/day):M: 1.7; W: 1.7 | Fifth quintile (cup eq/day):M: ≥3.0; W: ≥2.9 | ||
| Fruits, whole | ≥0.4 cup eq/1,000 kcal | ≥4 servings/day | |||
| Grains, whole | ≥1.5 ounce eq/1,000 kcal | M: 90; g/day W: 75 g/day | ≥ median (ounce eq/day):M: 0.9; W: 0.7 | Fifth quintile (ounce eq/day): M: ≥1.6; W: ≥1.3 | |
| Vegetables | ≥1.1 cup eq/1,000 kcal | ≥5 servings/day | ≥ median (cup eq/day):M: 1.3; W: 1.7 | Fifth quintile (cup eq/day):M: ≥2.7; W: ≥2.7 | |
| Greens and beans | ≥0.2 cup eq/1,000 kcal | ||||
| Total protein foods | ≥2.5 ounce eq/1,000 kcal | ||||
| Seafood and plant protein | ≥0.8 cup eq/1,000 kcal | ||||
| Fish | ≥ median (ounce eq/day):M: 0.5; W: 0.4 | ||||
| Legumes | ≥ median (cup eq/day):M: 0.1; W: 0.04 | ||||
| Nuts | ≥ median (ounce eq/day):M: 0.3; W: 0.2 | ||||
| Nuts and legumes | ≥1 serving/day | Fifth quintile (ounce eq/day) M: ≥1.0; W: ≥0.6 | |||
| Dairy | ≥1.3 cup eq/1,000 kcal | ||||
| Low-fat dairy | Fifth quintile (cup eq/day)M: ≥2.1; W: ≥1.9 | ||||
| Moderation components | |||||
| Total fat | ≤27% kcal/day | ||||
| Saturated fat | ≤8% kcal/day | ≤6% kcal/day | |||
| Trans-fat | ≤0.5% kcal/day | ||||
| Cholesterol | ≤71.4 mg/1,000 kcal | ||||
| Sodium | ≤1.1 g/1,000 kcal | Lowest decile (mg/day): M: ≤ 1,609; W: ≤ 1,242 | First quintile (mg/day): M: ≤ 1970; W: ≤ 1,531 | ≤1,143 mg/1,000 kcal | |
| Red and processed meats | 0 servings/day | < median (ounce eq/day): M: 2.19; W: 1.27 | First quintile (ounce eq/day): M: ≤ 1.1; W: ≤ 0.6 | ||
| SSB and fruit juices | 0 servings/day | First quintile (cup eq/day) M: ≤ 0.2; W: ≤ 0.1 | |||
| Alcohol | drinks/day: M: 0.5–2.0; W: 0.5–1.5 | M: 10–25; W: 5–15 g/day | |||
| Grains, refined | ≤1.8 ounce eq/1,000 kcal | ||||
| Added sugars | ≤6.5% kcal/day | ||||
Abbreviations: AHEI-2010, Alternative Healthy Eating Index-2010, aMed, alternate Mediterranean diet; DASH, Dietary Approaches to Stop Hypertension; DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; eq, equivalent; HEI-2015, Healthy Eating Index-2015; M, men; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; SSB, sugar-sweetened beverages; W, women.
a Sex specific medians.
b Sex-specific quintiles.
Cohort follow-up and case ascertainment
Cancer cases were identified by linking the cohort participants to 11 state registries (including the 8 states mentioned above plus Arizona, Nevada, and Texas) and the National Death Index from 1995 through 2011 (the final year for which linkage was performed). The cancer registries are estimated to be about 90% complete (35). Vital status was determined via linkage to the Social Security Administration Death Master File.
Our outcome was incident primary adenocarcinoma of the exocrine pancreas (International Classifications of Diseases for Oncology, Third Edition, codes C250 to C259). Our case definition excluded pancreatic endocrine tumors, sarcomas, and lymphomas (histology types 8150, 8151, 8153, 8155, and 8240), as their etiologies are thought to differ.
Statistical analysis
Spearman correlation coefficients were performed to assess the correspondence between the 5 dietary pattern scores. We calculated follow-up time from date of baseline questionnaire to PDAC diagnosis, death, move from study area, or end of follow-up (December 31, 2011), whichever occurred first. Cox proportional hazards models were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for PDAC. We used sex-specific quintiles to categorize each score, with the lowest quintile serving as the referent category and the highest quintile representing the highest diet quality. Continuous HRs (95% CIs) and P values for trend were based on a 1-standard-deviation increase in dietary quality score. We tested for confounding by the variables in Tables 2 and 3 beyond age and sex. A confounder was associated with both the dietary quality index and PDAC and changed the HRs by 10% or more (36). As none were confounders, all multivariable models adjusted for total energy intake (kcal/day) and putative risk factors for PDAC including for age at baseline (years, continuous), sex (for sex-combined analysis), smoking status (never smoker; quit >10 years ago, 5–9 years ago, 1–4 years ago, or <1 year ago; current smoker ≤20 cigarettes/day or >20 cigarettes/day; or missing), body mass index (calculated as weight (kg)/height (m)2: <25.0, 25.0–29.9, ≥30.0, or missing), and diabetes (yes vs. no). We tested for interactions by sex, race/ethnicity (non-Hispanic White, non-Hispanic Black, and others; others include Hispanics, Asian, Pacific Islander or American Indian, and Alaskan Native), smoking (never/former quit >10 years ago, smoker/quit <10 years ago), body mass index (<25, ≥25), and alcohol consumption (<3 drinks/day, ≥3 drinks/day) (6) using likelihood ratio tests comparing regression models with and without multiplicative term for the continuous score of each diet quality index. Wald tests were used to determine P values per 1-standard-deviation score increase and P for interaction. We evaluated the proportional hazards assumption by modeling the interaction term for the continuous score of each diet quality index and follow-up time. All proportionality tests showed P values of >0.05, meaning insufficient evidence for violations of proportional hazards. We conducted sensitivity analyses including only first primary PDAC and 5-year lagged analyses, excluding cases that developed PDAC within the first 5 years of follow-up by delaying the start of follow-up for all participants, to evaluate potential effects of reverse causation.
Table 2.
Baseline Characteristics According to Quintilea of Dietary Index for Men (n = 315,780), National Institutes of Health–AARP Diet and Health Study, United States 1995–2011
| HEI-2015 Quintiles | AHEI-2010 Quintile | aMed Quintile | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Characteristic | 1 | 3 | 5 | 1 | 3 | 5 | 1 | 3 | 5 | |||||||||
| Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | |
| Median score | 52.5 (5.1) | 67.7 (1.5) | 79.5 (3.2) | 38.5 (4.1) | 52.0 (1.5) | 66.4 (4.8) | 2.5 (0.7) | 5.0 (0.0) | 7.4 (0.6) | |||||||||
| Age at baseline, years | 61.5 (5.5) | 62.3 (5.3) | 63.1 (5.1) | 61.9 (5.5) | 62.3 (5.3) | 62.6 (5.2) | 62.1 (5.4) | 62.3 (5.3) | 62.5 (5.2) | |||||||||
| Self-reported raceb | ||||||||||||||||||
| Non-Hispanic White | 92.8 | 92.3 | 93.1 | 92.9 | 92.4 | 92.8 | 92.9 | 92.5 | 92.8 | |||||||||
| Non-Hispanic Black | 2.6 | 2.6 | 2.8 | 3.1 | 2.8 | 2.1 | 2.4 | 2.8 | 3.0 | |||||||||
| Others | 3.2 | 3.9 | 3.1 | 2.6 | 3.7 | 4.0 | 3.3 | 3.7 | 3.3 | |||||||||
| College graduate or postgraduate | 32.1 | 45.7 | 54.6 | 34.0 | 44.0 | 56.2 | 37.1 | 45.1 | 52.1 | |||||||||
| Body mass indexc | 27.5 (4.7) | 27.5 (4.3) | 26.6 (3.9) | 27.5 (4.5) | 27.4 (4.3) | 26.7 (4.1) | 27.4 (4.4) | 27.3 (4.3) | 27.0 (4.3) | |||||||||
| Smoking | ||||||||||||||||||
| Never smoker | 23.3 | 29.5 | 33.5 | 26.5 | 29.4 | 31.0 | 24.0 | 28.9 | 34.7 | |||||||||
| Quit ≥10 years ago | 35.5 | 45.6 | 49.1 | 37.5 | 44.3 | 49.6 | 40.5 | 44.7 | 46.2 | |||||||||
| Quit 1–9 years ago | 13.2 | 11.4 | 8.9 | 12.6 | 11.1 | 9.8 | 13.0 | 11.4 | 9.0 | |||||||||
| Current smoker or stopped within 1 year | 23.7 | 9.7 | 4.9 | 19.4 | 11.3 | 6.0 | 18.2 | 11.1 | 6.5 | |||||||||
| Self-reported diabetes | 9.1 | 11.0 | 10.1 | 9.1 | 10.9 | 9.9 | 9.3 | 10.9 | 10.2 | |||||||||
| Physical activity ≥20 minutes, ≥5×/week | 14.9 | 20.5 | 29.3 | 15.1 | 20.1 | 29.8 | 15.0 | 21.5 | 29.1 | |||||||||
| Regular multivitamin use | 44.3 | 52.4 | 58.7 | 46.5 | 52.1 | 57.0 | 46.5 | 52.1 | 57.9 | |||||||||
| Alcohol intake, ≥3 drinks/day | 9.8 | 11.4 | 10.2 | 20.6 | 10.1 | 3.0 | 20.1 | 9.4 | 3.2 | |||||||||
| Mean daily dietary intake | ||||||||||||||||||
| Energy, kcals/day | 2,243 (999) | 1,992 (816) | 1,836 (688) | 2,104 (903) | 1,987 (845) | 1,984 (774) | 1,730 (808) | 2,007 (798) | 2,365 (843) | |||||||||
| Red meat, ounce eq/1,000 kcal | 1.2 (0.7) | 1.0 (0.6) | 0.7 (0.4) | 1.1 (0.7) | 1.0 (0.6) | 0.7 (0.5) | 1.1 (0.7) | 0.9 (0.6) | 0.7 (0.5) | |||||||||
| Processed meat, ounce eq/1,000 kcal | 1.0 (1.0) | 0.7 (0.7) | 0.5 (0.5) | 1.0 (0.9) | 0.7 (0.7) | 0.4 (0.6) | 0.7 (0.7) | 0.7 (0.7) | 0.7 (0.7) | |||||||||
| Median score | 18.5 (1.6) | 24.0 (0.0) | 29.4 (1.6) | 1.1 (0.4) | 3.2 (0.2) | 5.8 (0.8) | ||||||||||||
| Age at baseline, years | 61.5 (5.5) | 62.4 (5.3) | 62.8 (5.2) | 61.8 (5.4) | 62.3 (5.3) | 62.8 (5.2) | ||||||||||||
| Self-reported raceb | ||||||||||||||||||
| Non-Hispanic White | 90.9 | 93.3 | 93.5 | 92.7 | 92.9 | 92.4 | ||||||||||||
| Non-Hispanic Black | 3.6 | 2.6 | 2.1 | 3.6 | 2.6 | 2.1 | ||||||||||||
| Others | 4.1 | 3.1 | 3.4 | 2.6 | 3.4 | 4.3 | ||||||||||||
| College graduate or postgraduate | 35.9 | 45.0 | 52.6 | 36.8 | 44.1 | 52.0 | ||||||||||||
| Body mass indexc | 27.5 (4.5) | 27.4 (4.3) | 24.7 (4.2) | 27.6 (4.6) | 27.5 (4.3) | 26.6 (4.0) | ||||||||||||
| Smoking | ||||||||||||||||||
| Never smoker | 24.0 | 28.9 | 34.5 | 26.4 | 28.0 | 32.8 | ||||||||||||
| Quit ≥10 years ago | 38.3 | 44.7 | 47.1 | 38.0 | 44.6 | 47.8 | ||||||||||||
| Quit 1–9 years ago | 13.5 | 11.2 | 8.9 | 12.3 | 11.5 | 9.9 | ||||||||||||
| Current smoker or stopped within 1 year | 20.0 | 11.3 | 5.8 | 19.6 | 12.0 | 5.5 | ||||||||||||
| Self-reported diabetes | 8.2 | 10.6 | 11.6 | 7.3 | 11.0 | 11.5 | ||||||||||||
| Physical activity ≥20 minutes, ≥5×/week | 14.1 | 20.0 | 31.4 | 15.9 | 19.4 | 30.1 | ||||||||||||
| Regular multivitamin use | 44.8 | 51.7 | 58.6 | 45.4 | 51.5 | 58.3 | ||||||||||||
| Alcohol intake, ≥3 drinks/day | 16.2 | 10.5 | 6.4 | 12.0 | 12.1 | 5.6 | ||||||||||||
| Mean daily dietary intake | ||||||||||||||||||
| Energy, kcals/day | 1,914 (847) | 1,989 (833) | 2,181 (829) | 2,231 (917) | 2,006 (817) | 1,745 (711) | ||||||||||||
| Red meat, ounce eq/1,000 kcal | 1.3 (0.7) | 0.9 (0.5) | 0.6 (0.4) | 1.2 (0.6) | 1.0 (0.6) | 0.6 (0.4) | ||||||||||||
| Processed meat, ounce eq/1,000 kcal | 0.9 (0.9) | 0.7 (0.7) | 0.5 (0.6) | 1.0 (0.9) | 0.7 (0.7) | 0.5 (0.6) | ||||||||||||
Abbreviations: AHEI-2010, Alternative Heathy Eating Index; aMed, alternate Mediterranean diet; DASH, Dietary Approaches to Stop Hypertension; eq, equivalent; HEI-2015, Healthy Eating Index-2015; SD, standard deviation.
a Quintile 1, lowest; quintile 3, middle; quintile 5, highest.
b Other races/ethnicities include Hispanic, Asian, Pacific Islander or American Indian, and Alaskan Native.
c Weight (kg)/height (m)2.
Table 3.
Baseline Characteristics According to Quintilea of Dietary Index for Women (n = 220,044), National Institutes of Health–AARP Diet and Health Study, United States, 1995–2011
| HEI-2015 Quintile | AHEI-2010 Quintile | aMed Quintile | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 3 | 5 | 1 | 3 | 5 | 1 | 3 | 5 | ||||||||||
| Characteristic | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % |
| Median score | 54.6 (5.4) | 69.7 (1.4) | 80.6 (3.0) | 39.8 (3.9) | 52.4 (1.4) | 66.2 (4.5) | 2.6 (0.6) | 5.0 (0.0) | 7.4 (0.6) | |||||||||
| Age at baseline, years | 61.3 (5.5) | 61.9 (5.4) | 62.7 (5.2) | 61.7 (5.5) | 62.0 (5.4) | 62.0 (5.4) | 61.8 (5.5) | 61.9 (5.4) | 62.1 (5.3) | |||||||||
| Self-reported raceb | ||||||||||||||||||
| Non-Hispanic White | 90.6 | 89.5 | 89.3 | 89.4 | 89.4 | 90.5 | 91.6 | 90.0 | 87.2 | |||||||||
| Non-Hispanic Black | 4.9 | 5.3 | 6.1 | 6.2 | 5.5 | 4.4 | 3.8 | 5.2 | 7.4 | |||||||||
| Others | 2.8 | 3.7 | 3.3 | 2.6 | 3.5 | 3.7 | 2.8 | 3.4 | 3.9 | |||||||||
| College graduate or postgraduate | 21.1 | 30.7 | 37.1 | 21.7 | 29.1 | 39.9 | 24.3 | 30.1 | 35.8 | |||||||||
| Body mass indexc | 27.6 (6.7) | 26.9 (5.9) | 26.0 (5.5) | 27.4 (6.4) | 27.1 (6.3) | 25.9 (5.7) | 26.9 (6.1) | 26.9 (6.0) | 26.7 (6.0) | |||||||||
| Smoking | ||||||||||||||||||
| Never smoker | 37.9 | 44.5 | 47.0 | 44.5 | 44.6 | 40.2 | 38.4 | 43.6 | 48.4 | |||||||||
| Quit ≥10 years ago | 19.4 | 26.5 | 30.8 | 19.5 | 25.1 | 33.3 | 22.4 | 26.3 | 28.7 | |||||||||
| Quit 1–9 years ago | 11.2 | 11.0 | 10.3 | 10.2 | 11.3 | 11.8 | 11.7 | 11.4 | 9.5 | |||||||||
| Current smoker or stopped within 1 year | 27.9 | 14.4 | 8.6 | 22.4 | 15.6 | 11.0 | 23.8 | 15.4 | 9.9 | |||||||||
| Self-reported diabetes | 7.3 | 7.6 | 7.3 | 7.8 | 7.6 | 6.6 | 6.9 | 7.6 | 7.7 | |||||||||
| Physical activity ≥20 minutes, ≥5×/week | 10.2 | 15.8 | 22.7 | 10.5 | 15.2 | 23.9 | 11.2 | 15.9 | 21.6 | |||||||||
| Regular multivitamin use | 52.5 | 61.4 | 66.2 | 55.4 | 60.6 | 65.1 | 55.0 | 60.5 | 66.9 | |||||||||
| Alcohol intake, ≥3 drinks/day | 2.8 | 2.8 | 2.4 | 5.9 | 2.3 | 0.8 | 6.0 | 2.3 | 0.5 | |||||||||
| Mean daily dietary intake | ||||||||||||||||||
| Energy, kcals/day | 1,686 (758) | 1,155 (643) | 1,485 (565) | 1,565 (647) | 1,544 (665) | 1,643 (655) | 1,225 (519) | 1,560 (596) | 1,960 (695) | |||||||||
| Red meat, ounce eq/1,000 kcal | 1.0 (0.6) | 0.8 (0.5) | 0.5 (0.4) | 1.0 (0.6) | 0.8 (0.5) | 0.5 (0.4) | 1.0 (0.6) | 0.8 (0.5) | 0.6 (0.4) | |||||||||
| Processed meat, ounce eq/1,000 kcal | 0.5 (0.6) | 0.4 (0.5) | 0.3 (0.3) | 0.5 (0.6) | 0.4 (0.5) | 0.3 (0.4) | 0.4 (0.5) | 0.4 (0.5) | 0.4 (0.5) | |||||||||
| Median score | 18.5 (1.2) | 23.5 (0.5) | 29.4 (1.6) | 1.4 (0.6) | 3.7 (0.2) | 6.7 (0.6) | ||||||||||||
| Age at baseline, years | 61.4 (5.4) | 62.0 (5.4) | 62.4 (5.3) | 61.5 (5.4) | 61.9 (5.4) | 62.5 (5.3) | ||||||||||||
| Self-reported raceb | ||||||||||||||||||
| Non-Hispanic White | 87.8 | 90.0 | 90.5 | 89.1 | 89.5 | 90.6 | ||||||||||||
| Non-Hispanic Black | 7.0 | 5.2 | 4.6 | 6.7 | 5.6 | 4.0 | ||||||||||||
| Others | 3.4 | 3.3 | 3.5 | 2.8 | 3.3 | 3.8 | ||||||||||||
| College graduate or postgraduate | 22.4 | 29.2 | 38.0 | 23.1 | 29.4 | 37.4 | ||||||||||||
| Body mass indexc | 27.4 (6.4) | 27.1 (6.1) | 26.1 (5.8) | 27.7 (6.6) | 27.0 (6.0) | 25.8 (5.3) | ||||||||||||
| Smoking | ||||||||||||||||||
| Never smoker | 38.6 | 43.5 | 47.1 | 40.6 | 43.7 | 45.4 | ||||||||||||
| Quit ≥10 years ago | 20.3 | 25.9 | 31.0 | 20.8 | 25.8 | 30.7 | ||||||||||||
| Quit 1–9 years ago | 11.6 | 11.3 | 9.8 | 11.0 | 11.2 | 11.0 | ||||||||||||
| Current smoker or stopped within 1 year | 25.8 | 15.7 | 8.7 | 24.3 | 15.7 | 9.1 | ||||||||||||
| Self-reported diabetes | 7.1 | 7.7 | 7.7 | 6.8 | 8.0 | 7.1 | ||||||||||||
| Physical activity ≥20 minutes, ≥5×/week | 9.7 | 14.6 | 24.6 | 10.1 | 14.8 | 25.2 | ||||||||||||
| Regular multivitamin use | 52.4 | 60.4 | 66.8 | 54.0 | 61.0 | 66.0 | ||||||||||||
| Alcohol intake, ≥3 drinks/day | 5.1 | 2.7 | 1.2 | 3.7 | 3.8 | 0.5 | ||||||||||||
| Mean daily dietary intake | ||||||||||||||||||
| Energy, kcals/day | 1,393 (602) | 1,528 (629) | 1,791 (683) | 1,791 (741) | 1,570 (648) | 1,389 (557) | ||||||||||||
| Red meat, ounce eq/1,000 kcal | 1.1 (0.6) | 0.8 (0.5) | 0.4 (0.3) | 1.0 (0.6) | 0.8 (0.5) | 0.4 (0.3) | ||||||||||||
| Processed meat, ounce eq/1,000 kcal | 0.5 (0.6) | 0.4 (0.5) | 0.2 (0.4) | 0.5 (0.6) | 0.4 (0.5) | 0.2 (0.3) | ||||||||||||
Abbreviations: AHEI-2010, Alternative Heathy Eating Index; aMed, alternate Mediterranean diet; DASH, Dietary Approaches to Stop Hypertension; eq, equivalent; HEI-2015, Healthy Eating Index-2015.
a Quintile 1, lowest; quintile 3, middle; quintile 5, highest.
b Other races/ethnicities include Hispanic, Asian, Pacific Islander or American Indian, and Alaskan Native.
c Weight (kg)/height (m)2.
To determine whether an association of adherence to diet indices and PDAC was mediated by a specific food or nutrient (37), exploratory analyses were conducted to examine independent associations for individual components. Separate HRs (95% CI) were calculated for each component (component i) with adjustment for modified scores that did not include the respective components as follows:
Modified score = total score – component i (38).
P values of <0.05 were considered statistically significant; however, to account for multiple comparisons across the 15 associations with the 5 diet quality indices (sex-combined, men, and women), we note associations that were significant below the Bonferroni-corrected P value of <0.003 (0.05/15). All statistical analyses were performed with SAS (version 9.4; SAS Institute, Inc., Cary, North Carolina), and statistical tests are 2-sided.
RESULTS
During up to 16 years of follow-up (median 15.5 years), 3,137 (1,988 men and 1,149 women) incident PDAC cases were identified. Sex-specific selected baseline characteristics by diet quality are shown in Table 2 for men and Table 3 for women. Across all indices in both men and women, those in the highest- compared with lowest-quintiles were more likely to be slightly older, a college graduate or postgraduate, leaner, physically active, multivitamin users, never or former smokers having quit ≥10 years ago, and less likely to consume ≥3 drinks/day (except for HEI-2015) (Tables 2 and 3). All the diet quality indices were correlated (P < 0.0001; Web Table 1), with the strongest correlations between AHEI-2010 and DASH-Fung (r = 0.65).
In sex-combined multivariable-adjusted models, participants with the highest diet quality compared with those with the lowest (quintile 5 vs. quintile 1) had significantly lower PDAC risk (for HEI-2015, HR = 0.84, 95% CI: 0.75, 0.94, P for trend = <0.0001; aMed, HR = 0.82, 95% CI: 0.73, 0.93, P for trend = 0.0004; DASH-Fung, HR = 0.85, 95% CI: 0.77, 0.95, P for trend = 0.004; and DASH-Mellen, HR = 0.86, 95% CI: 0.77, 0.96, P for trend = 0.006), except for AHEI-2010: HR = 0.93, 95% CI: 0.83, 1.04 (Table 4). Similar patterns were observed for continuous diet quality indices—per 1-standard-deviation increase for HEI-2015 (HR = 0.99, 95% CI: 0.99, 1.00), aMed (HR = 0.96, 95% CI: 0.94, 0.98), DASH-Fung (HR = 0.99, 95% CI: 0.98, 1.00), and DASH-Mellen (HR = 0.97, 95% CI: 0.95, 0.99). HEI-2015 and aMed diet quality indices remained statistically significant with PDAC risk below the Bonferroni-corrected P value of <0.003. As a sensitivity analysis to evaluate potentially unmeasured confounding, we calculated the E-value for the continuous scores of diet quality indices for the significant associations in sex combined analyses (39, 40). The E-value represents the minimum association in terms of relative risk that an unmeasured confounder would need to have per 1-standard-deviation increase of the diet quality index with PDAC to fully explain the observed association (39, 40). The calculated E-values are 1.10 for HEI-2015, 1.24 for aMed, 1.13 for DASH-Fung, and 1.21 for DASH-Mellen. The small E-values suggest small unmeasured confounding could explain our observed associations.
Table 4.
Sex-Combined Hazard Ratios for Pancreatic Ductal Adenocarcinoma According to Quintile of Dietary Index (n = 535,824) in the National Institutes of Health–AARP Diet and Health Study, United States, 1995–2011
| Diet Quality | No. | Person-Years | No. ofPDACCases | Age- and Sex-AdjustedHR a | 95%CI | P forTrend b | Multivariable-AdjustedHR c | 95%CI | P forTrend b |
|---|---|---|---|---|---|---|---|---|---|
| HEI-2015 | |||||||||
| Quintile 1 (lowest) | 107,165 | 1,408,256 | 649 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 107,165 | 1,429,681 | 657 | 0.96 | 0.86, 1.07 | 0.97 | 0.87, 1.09 | ||
| Quintile 3 | 107,165 | 1,440,965 | 629 | 0.89 | 0.80, 0.99 | 0.91 | 0.82, 1.02 | ||
| Quintile 4 | 107,165 | 1,449,913 | 598 | 0.82 | 0.74, 0.92 | 0.85 | 0.76, 0.95 | ||
| Quintile 5 (highest) | 107,164 | 1,461,910 | 604 | 0.80 | 0.71, 0.89 | 0.84 | 0.75, 0.94 | ||
| Continuousb | 0.99 | 0.99, 0.99 | <0.0001 | 0.99 | 0.99, 1.00 | <0.0001 | |||
| AHEI-2010 | |||||||||
| Quintile 1 (lowest) | 107,165 | 1,408,789 | 628 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 107,165 | 1,426,679 | 607 | 0.93 | 0.83, 1.04 | 0.94 | 0.84, 1.05 | ||
| Quintile 3 | 107,165 | 1,439,012 | 633 | 0.95 | 0.85, 1.06 | 0.96 | 0.86, 1.04 | ||
| Quintile 4 | 107,165 | 1,448,879 | 641 | 0.95 | 0.85, 1.06 | 0.96 | 0.86, 1.08 | ||
| Quintile 5 (highest) | 107,164 | 1,467,366 | 628 | 0.91 | 0.82, 1.02 | 0.93 | 0.83, 1.04 | ||
| Continuousb | 1.00 | 0.99, 1.00 | 0.11 | 1.00 | 0.99, 1.00 | 0.25 | |||
| aMed | |||||||||
| Quintile 1 (lowest) | 121,940 | 1,605,797 | 752 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 101,616 | 1,357,028 | 632 | 0.98 | 0.88, 1.09 | 0.98 | 0.88, 1.09 | ||
| Quintile 3 | 110,164 | 1,480,186 | 627 | 0.89 | 0.80, 0.99 | 0.88 | 0.79, 0.98 | ||
| Quintile 4 | 126,137 | 1,318,288 | 548 | 0.86 | 0.77, 0.96 | 0.85 | 0.77, 0.95 | ||
| Quintile 5 (highest) | 75,967 | 1,429,426 | 578 | 0.83 | 0.75, 0.93 | 0.82 | 0.73, 0.93 | ||
| Continuousb | 0.96 | 0.95, 0.98 | 0.003 | 0.96 | 0.94, 0.98 | <0.0001 | |||
| DASH-Fung | |||||||||
| Quintile 1 (lowest) | 137,469 | 1,821,046 | 836 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 102,388 | 1,370,756 | 585 | 0.89 | 0.80, 0.99 | 0.90 | 0.81, 1.00 | ||
| Quintile 3 | 110,007 | 1,478,219 | 644 | 0.90 | 0.81, 0.99 | 0.90 | 0.81, 1.00 | ||
| Quintile 4 | 92,204 | 1,244,921 | 534 | 0.87 | 0.78, 0.97 | 0.88 | 0.79, 0.98 | ||
| Quintile 5 (highest) | 93,756 | 1,275,784 | 538 | 0.84 | 0.76, 0.94 | 0.85 | 0.77, 0.95 | ||
| Continuousb | 0.98 | 0.98, 0.99 | 0.001 | 0.99 | 0.98, 1.00 | 0.004 | |||
| DASH-Mellen | |||||||||
| Quintile 1 (lowest) | 145,245 | 1,926,722 | 895 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 112,929 | 1,508,785 | 663 | 0.92 | 0.84, 1.02 | 0.92 | 0.83, 1.02 | ||
| Quintile 3 | 111,671 | 1,501,068 | 640 | 0.88 | 0.80, 0.97 | 0.89 | 0.80, 0.99 | ||
| Quintile 4 | 81,945 | 1,111,486 | 468 | 0.86 | 0.77, 0.96 | 0.88 | 0.79, 0.99 | ||
| Quintile 5 (highest) | 84,034 | 1,142,666 | 471 | 0.83 | 0.74, 0.93 | 0.86 | 0.77, 0.96 | ||
| Continuousb | 0.96 | 0.95, 0.98 | <0.0001 | 0.97 | 0.95, 0.99 | 0.006 |
Abbreviations: AHEI-2010, Alternative Heathy Eating Index; aMed, alternate Mediterranean diet; CI, confidence interval; DASH, Dietary Approaches to Stop Hypertension; HEI-2015, Healthy Eating Index-2015; HR, hazard ratio; PDAC, pancreatic ductal adenocarcinoma.
a Estimated using Cox proportional hazard regression model with person-years as the underlying time metric. HRs compares the risk of developing PDAC for participants in each quintile of diet quality score compared with participants in the lowest quintile (lower adherence).
b HRs (95% CIs) and P for trend per 1-standard deviation increase.
c Multivariable models adjusted for age at baseline (years, continuous), sex, smoking status (never smoker, quit >10 years ago, quit 5–9 years ago, quit 1–4 years ago, quit <1 year or current smoker ≤20 cigarettes/day, quit <1 year or current smoker >20 cigarettes/day, or missing), body mass index (weight (kg)/height (m)2: <25.0, 25.0–29.9, ≥30.0, or missing), diabetes (yes vs. no), and total energy intake (kcal/day).
Although interaction by sex was not statistically significant (P for interaction > 0.07 for all indices), the pattern of associations for adherence to 4 of the diet quality indices differed by sex (Tables 5 and 6). In men, the highest- compared with the lowest-quintile diet quality scores were statistically significantly associated with a lower PDAC risk for HEI-2015 (HR = 0.78, 95% CI: 0.68, 0.90), aMed (HR = 0.85, 95% CI: 0.74, 0.98), DASH-Fung (HR = 0.77, 95% CI: 0.66, 0.90), and DASH-Mellen (HR = 0.82, 95% CI: 0.71, 0.95), except for AHEI-2010 (HR = 0.89, 95% CI: 0.77, 1.02). Similar patterns were observed for continuous dietary pattern scores per 1-standard-deviation increase (for HEI-2015, HR = 0.99, 95% CI: 0.99, 1.00, P for trend < 0.0001; aMed, HR = 0.97, 95% CI: 0.95, 1.00, P for trend = 0.04; DASH-Fung, HR = 0.98, 95% CI: 0.97, 0.99, P for trend = 0.002; and DASH-Mellen: HR = 0.95, 95% CI: 0.93, 0.98, P for trend = 0.0006). The HEI-2015, DASH-Fung, and DASH-Mellen were significantly associated with PDAC risk below the Bonferroni threshold. In women, only aMed diet quality showed a statistically significant association with PDAC, with those in the highest quintile having a lower risk (HR = 0.76, 95% CI: 0.63, 0.92). Similarly, when evaluating the aMed score as a continuous measure, scores were inversely associated with PDAC (HR = 0.94, 95% CI: 0.91, 0.98; P for trend = 0.001), which remained statistically significant after Bonferroni correction.
Table 5.
Hazard Ratios for Pancreatic Ductal Adenocarcinoma According to Quintile of Dietary Index for Men (n = 315,780) in the National Institutes of Health–AARP Diet and Health Study, United States, 1995–2011
| Diet Quality | No. | Person-Years | No. ofPDAC Cases | Age-AdjustedHR a | 95%CI | P forTrend b | Multivariable-Adjusted HR c | 95%CI | P forTrend b |
|---|---|---|---|---|---|---|---|---|---|
| HEI-2015 | |||||||||
| Quintile 1 (lowest) | 63,156 | 816,953 | 421 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 63,156 | 828,856 | 422 | 0.95 | 0.83, 1.09 | 0.96 | 0.83, 1.09 | ||
| Quintile 3 | 63,156 | 836,086 | 391 | 0.85 | 0.74, 0.98 | 0.86 | 0.75, 0.99 | ||
| Quintile 4 | 63,156 | 841,643 | 381 | 0.80 | 0.70, 0.92 | 0.82 | 0.72, 0.95 | ||
| Quintile 5 (highest) | 63,156 | 849,410 | 373 | 0.75 | 0.66, 0.87 | 0.78 | 0.68, 0.90 | ||
| Continuousb | 0.99 | 0.99, 0.99 | <0.0001 | 0.99 | 0.99, 1.00 | <0.0001 | |||
| AHEI-2010 | |||||||||
| Quintile 1 (lowest) | 63,156 | 816,127 | 404 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 63,156 | 826,876 | 390 | 0.93 | 0.81, 1.06 | 0.93 | 0.81, 1.07 | ||
| Quintile 3 | 63,156 | 833,982 | 405 | 0.94 | 0.82, 1.08 | 0.95 | 0.83, 1.09 | ||
| Quintile 4 | 63,156 | 841,860 | 397 | 0.91 | 0.79, 1.04 | 0.92 | 0.80, 1.05 | ||
| Quintile 5 (highest) | 63,156 | 854,104 | 392 | 0.87 | 0.76, 1.00 | 0.89 | 0.77, 1.02 | ||
| Continuousb | 1.00 | 0.99, 1.00 | 0.03 | 1.00 | 0.99, 1.00 | 0.07 | |||
| aMed | |||||||||
| Quintile 1 (lowest) | 74,653 | 967,310 | 480 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 59,913 | 787,764 | 398 | 1.00 | 0.88, 1.14 | 1.00 | 0.88, 1.14 | ||
| Quintile 3 | 64,601 | 854,540 | 391 | 0.90 | 0.79, 1.03 | 0.90 | 0.78, 1.02 | ||
| Quintile 4 | 56,399 | 753,150 | 360 | 0.94 | 0.82, 1.07 | 0.93 | 0.81, 1.07 | ||
| Quintile 5 (highest) | 60,214 | 810,185 | 359 | 0.86 | 0.75, 0.98 | 0.85 | 0.74, 0.98 | ||
| Continuousb | 0.97 | 0.95, 1.00 | 0.03 | 0.97 | 0.95, 1.00 | 0.04 | |||
| DASH-Fung | |||||||||
| Quintile 1 (lowest) | 55,786 | 727,251 | 376 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 85,120 | 1,119,633 | 531 | 0.88 | 0.77, 1.00 | 0.88 | 0.77, 1.00 | ||
| Quintile 3 | 32,847 | 434,230 | 214 | 0.89 | 0.76, 1.06 | 0.89 | 0.76, 1.06 | ||
| Quintile 4 | 86,892 | 1,152,784 | 543 | 0.84 | 0.74, 0.96 | 0.84 | 0.74, 0.96 | ||
| Quintile 5 (highest) | 55,135 | 739,051 | 324 | 0.77 | 0.66, 0.89 | 0.77 | 0.66, 0.90 | ||
| Continuousb | 0.98 | 0.97, 0.99 | 0.0009 | 0.98 | 0.97, 0.99 | 0.002 | |||
| DASH-Mellen | |||||||||
| Quintile 1 (lowest) | 55,145 | 720,680 | 360 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 61,838 | 811,813 | 431 | 1.04 | 0.91, 1.20 | 1.04 | 0.90, 1.19 | ||
| Quintile 3 | 71,040 | 937,207 | 432 | 0.89 | 0.77, 1.02 | 0.89 | 0.77, 1.02 | ||
| Quintile 4 | 59,862 | 794,981 | 371 | 0.89 | 0.77, 1.03 | 0.90 | 0.77, 1.04 | ||
| Quintile 5 (highest) | 67,895 | 908,269 | 394 | 0.81 | 0.70, 0.93 | 0.82 | 0.71, 0.95 | ||
| Continuousb | 0.95 | 0.93, 0.98 | 0.0002 | 0.95 | 0.93, 0.98 | 0.0006 |
Abbreviations: AHEI-2010, Alternative Healthy Eating Index-2010; aMed, alternate Mediterranean diet; CI, confidence interval; DASH, Dietary Approaches to Stop Hypertension; HEI-2015, Healthy Eating Index-2015; HR, hazard ratio; PDAC, pancreatic ductal adenocarcinoma.
a Estimated using Cox proportional hazard regression model with person-years as the underlying time metric. HRs compares the risk of developing PDAC for participants in each quintile of diet quality score compared with participants in the lowest quintile (lower adherence).
b Hazard ratios and P for trend per 1-standard deviation increase. P for interaction by sex > 0.07 for all scores.
c Multivariable models adjusted for age at baseline (years, continuous), smoking status (never smoker, quit >10 years ago, quit 5–9 years ago, quit 1–4 years ago, quit <1 year or current smoker ≤20 cigarettes/day, quit <1 year or current smoker >20 cigarettes/day, or missing), body mass index (weight (kg)/height (m)2: <25.0, 25.0–29.9, ≥30.0, or missing), diabetes (yes vs. no), and total energy intake (kcal/day).
Table 6.
Hazard Ratios for Pancreatic Ductal Adenocarcinoma According to Quintile of Dietary Index for Women (n = 220,044) in the National Institutes of Health–AARP Diet and Health Study, United States, 1995–2011
| Diet Quality | No. | Person-Years | No. of PDACCases | Age-AdjustedHR a | 95%CI | P forTrend b | Multivariable-Adjusted HR c | 95% CI | P forTrend b |
|---|---|---|---|---|---|---|---|---|---|
| HEI-2015 | |||||||||
| Quintile 1 (lowest) | 44,008 | 591,288 | 228 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 44,009 | 600,825 | 235 | 0.98 | 0.82, 1.17 | 1.01 | 0.84, 1.21 | ||
| Quintile 3 | 44,009 | 604,882 | 238 | 0.97 | 0.81, 1.16 | 1.02 | 0.85, 1.22 | ||
| Quintile 4 | 44,009 | 608,267 | 217 | 0.85 | 0.71, 1.03 | 0.91 | 0.75, 1.10 | ||
| Quintile 5 (highest) | 44,009 | 612,515 | 231 | 0.87 | 0.73, 1.05 | 0.94 | 0.78, 1.13 | ||
| Continuousb | 0.99 | 0.99, 1.00 | 0.03 | 1.00 | 0.99, 1.00 | 0.19 | |||
| AHEI-2010 | |||||||||
| Quintile 1 (lowest) | 44,008 | 592,646 | 224 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 44,009 | 599,802 | 217 | 0.94 | 0.78, 1.13 | 0.95 | 0.79, 1.14 | ||
| Quintile 3 | 44,009 | 605,031 | 228 | 0.97 | 0.81, 1.17 | 0.98 | 0.82, 1.18 | ||
| Quintile 4 | 44,009 | 607,019 | 244 | 1.03 | 0.86, 1.24 | 1.04 | 0.87, 1.25 | ||
| Quintile 5 (highest) | 44,009 | 613,278 | 236 | 0.99 | 0.82, 1.19 | 1.00 | 0.83, 1.20 | ||
| Continuousb | 1.00 | 1.00, 1.01 | 0.72 | 1.00 | 1.00, 1.01 | 0.67 | |||
| aMed | |||||||||
| Quintile 1 (lowest) | 47,287 | 638,487 | 272 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 41,703 | 569,264 | 234 | 0.95 | 0.80, 1.13 | 0.94 | 0.79, 1.12 | ||
| Quintile 3 | 45,563 | 625,646 | 236 | 0.87 | 0.73, 1.04 | 0.85 | 0.71, 1.02 | ||
| Quintile 4 | 40,898 | 565,138 | 188 | 0.76 | 0.63, 0.92 | 0.73 | 0.61, 0.89 | ||
| Quintile 5 (highest) | 44,593 | 619,241 | 219 | 0.80 | 0.67, 0.96 | 0.76 | 0.63, 0.92 | ||
| Continuousb | 0.95 | 0.92, 0.98 | 0.002 | 0.94 | 0.91, 0.98 | 0.001 | |||
| DASH-Fung | |||||||||
| Quintile 1 (lowest) | 39,796 | 536,336 | 206 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 37,296 | 509,020 | 200 | 0.99 | 0.81, 1.20 | 1.00 | 0.82, 1.22 | ||
| Quintile 3 | 44,623 | 612,127 | 218 | 0.88 | 0.73, 1.07 | 0.90 | 0.74, 1.08 | ||
| Quintile 4 | 59,708 | 823,561 | 311 | 0.92 | 0.77, 1.10 | 0.94 | 0.79, 1.12 | ||
| Quintile 5 (highest) | 38,621 | 536,732 | 214 | 0.96 | 0.79, 1.16 | 0.98 | 0.80, 1.19 | ||
| Continuousb | 0.99 | 0.98, 1.01 | 0.32 | 0.99 | 0.98, 1.01 | 0.47 | |||
| DASH-Mellen | |||||||||
| Quintile 1 (lowest) | 42,346 | 571,385 | 215 | 1.00 | Referent | 1.00 | Referent | ||
| Quintile 2 | 40,317 | 548,363 | 216 | 1.02 | 0.85, 1.23 | 1.04 | 0.86, 1.26 | ||
| Quintile 3 | 45,248 | 622,019 | 230 | 0.95 | 0.79, 1.14 | 0.98 | 0.82, 1.19 | ||
| Quintile 4 | 54,535 | 754,910 | 270 | 0.90 | 0.76, 1.08 | 0.96 | 0.80, 1.15 | ||
| Quintile 5 (highest) | 37,598 | 521,099 | 218 | 1.03 | 0.85, 1.24 | 1.10 | 0.91, 1.34 | ||
| Continuousb | 0.99 | 0.95, 1.02 | 0.38 | 1.00 | 0.97, 1.03 | 0.92 |
Abbreviations: AHEI-2010, Alternative Healthy Eating Index-2010; aMed, alternate Mediterranean diet; CI, confidence interval; DASH, Dietary Approaches to Stop Hypertension; HEI-2015, Healthy Eating Index-2015; HR, hazard ratio; PDAC, pancreatic ductal adenocarcinoma.
a Estimated using Cox proportional hazard regression model with person-years as the underlying time metric. HRs compares the risk of developing PDAC for participants in each quintile of diet quality score compared with participants in the lowest quintile (lower adherence).
b HRs (95% CI) and P for trend per 1-standard deviation increase. P for interaction by sex > 0.07 for all scores.
c Multivariable models adjusted for age at baseline (years, continuous), smoking status (never smoker, quit >10 years ago, quit 5–9 years ago, quit 1–4 years ago, quit <1 year or current smoker ≤20 cigarettes/day, quit <1 year or current smoker >20 cigarettes/day, or missing), body mass index (weight (kg)/height (m)2: <25.0, 25.0–29.9, ≥30.0, and missing), diabetes (yes vs. no), and total energy intake (kcal/day).
Table 7 shows the results for the exploratory analyses of components separately for each dietary pattern score. With the HEI-2015, greater alignment with the diet quality index recommendations (for whole grains (HR = 0.98, 95% CI: 0.96, 0.99), dairy (HR = 0.99, 95% CI: 0.98, 1.00), and saturated fat (HR = 0.99, 95% CI: 0.97, 1.00) were inversely associated with risk of PDAC, while added sugars (HR = 1.02, 95% CI: 1.00, 1.03) were positively associated with PDAC risk (P < 0.05). With the aMed, more optimal alignment with the recommendations for red and processed meat (HR = 0.92, 95% CI: 0.85, 0.99) and alcohol (HR = 0.87, 95% CI: 0.79, 0.96) consumption was associated with lower risk (P < 0.05). With the DASH-Fung, greater alignment with recommendations for total fruits (HR = 0.97, 95% CI: 0.95, 1.00) and whole grains (HR = 0.96, 95% CI: 0.94, 0.99) were inversely associated, whereas the sweetened beverages consumption component (HR = 1.04, 95% CI: 1.01, 1.06) was positively associated with PDAC risk (P < 0.05). Last, with the DASH-Mellen, more optimal alignment with calcium was associated with reduced risk (HR = 0.89, 95% CI: 0.80, 0.99).
Table 7.
Multivariable-Adjusted Hazard Ratios for Pancreatic Ductal Adenocarcinoma for Individuals Components of Dietary Index for Men and Women Combined (n = 535,824), National Institutes of Health–AARP Diet and Health Study, United States, 1995–2011
| Diet Quality Index | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Diet Quality Component | HEI-2015 | AHEI-2010 | aMed | DASH-Fung | DASH-Mellen | |||||
| HR a , b | 95% CI | HR a , b | 95% CI | HR a , b | 95% CI | HR a , b | 95% CI | HR a , b | 95% CI | |
| Adequacy Components | ||||||||||
| Calcium | 0.89 | 0.80, 0.99c | ||||||||
| Fiber | 0.90 | 0.79, 1.01 | ||||||||
| Magnesium | 0.99 | 0.85, 1.16 | ||||||||
| Potassium | 1.03 | 0.90, 1.18 | ||||||||
| Protein | 1.07 | 0.98, 1.17 | ||||||||
| MUFA: saturated fat | 0.93 | 0.87, 1.00 | ||||||||
| PUFA | 1.01 | 1.00, 1.02 | ||||||||
| PUFA+MUFA: saturated fat | 0.99 | 0.98, 1.00 | ||||||||
| EPA+DHA | 1.00 | 0.98, 1.01 | ||||||||
| Fruits, total | 0.97 | 0.94, 1.00 | 0.94 | 0.87, 1.01 | 0.97 | 0.95, 1.00c | ||||
| Fruits, whole | 0.97 | 0.94, 1.00 | 0.99 | 0.98, 1.00 | ||||||
| Grains, whole | 0.98 | 0.96, 0.99c | 0.95 | 0.93, 0.98c | 0.94 | 0.88, 1.02 | 0.96 | 0.94, 0.99c | ||
| Vegetables | 1.02 | 0.98, 1.05 | 0.99 | 0.98, 1.01 | 1.00 | 0.93, 1.08 | 0.99 | 0.96, 1.02 | ||
| Greens and beans | 0.99 | 0.97, 1.01 | ||||||||
| Total protein foods | 1.02 | 0.98, 1.06 | ||||||||
| Seafood and plant protein | 0.99 | 0.96, 1.02 | ||||||||
| Fish | 1.07 | 1.00, 1.16 | ||||||||
| Legumes | 0.96 | 0.89, 1.03 | ||||||||
| Nuts | 1.00 | 0.93, 1.07 | ||||||||
| Nuts and legumes | 0.99 | 0.98, 1.00c | 0.98 | 0.95, 1.00 | ||||||
| Dairy | 0.99 | 0.98, 1.00c | ||||||||
| Low-fat dairy | 0.98 | 0.95, 1.00 | ||||||||
| Moderation Components | ||||||||||
| Total fat | 1.01 | 0.92, 1.12 | ||||||||
| Saturated fat | 0.99 | 0.97, 1.00c | 0.97 | 0.84, 1.11 | ||||||
| Trans-fat | 1.01 | 1.00, 1.03 | ||||||||
| Cholesterol | 0.91 | 0.82, 1.01 | ||||||||
| Sodium | 1.00 | 0.98, 1.01 | 1.01 | 0.99, 1.03 | 1.01 | 0.97, 1.05 | 0.96 | 0.81, 1.13 | ||
| Red and processed meat | 0.99 | 0.97, 1.00 | 0.92 | 0.85, 0.99c | 0.98 | 0.95, 1.01 | ||||
| SSB and fruit juices | 1.01 | 1.00, 1.02c | 1.04 | 1.01, 1.06c | ||||||
| Alcohol | 0.99 | 0.98, 1.00 | 0.87 | 0.79, 0.96c | ||||||
| Grains, refined | 1.01 | 0.99, 1.02 | ||||||||
| Added sugars | 1.02 | 1.00, 1.03c | ||||||||
Abbreviations: AHEI-2010, Alternative Healthy Eating Index-2010; aMed, alternate Mediterranean diet; CI, confidence interval; DASH, Dietary Approaches to Stop Hypertension; DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; HEI-2015, Healthy Eating Index-2015; HR, hazard ratio; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; SSB, sugar-sweetened beverages.
a The HR is based on a 1-unit change in the score for the component of interest (meeting the recommendations vs. not meeting the recommendations).
b Multivariable models mutually adjusted for components by each score and also adjusted by age at baseline (years, continuous), sex (for sex-combined analysis), smoking status (never smoker, quit >10 years ago, quit 5–9 years ago, quit 1–4 years ago, quit < 1 year or current smoker ≤ 20 cigarettes/day, quit < 1 year or current smoker > 20 cigarettes/day, or missing), body mass index (weight (kg)/height (m)2: <25.0, 25.0–29.9, ≥30.0, or missing), diabetes (yes vs. no), and total energy intake (kcal/day).
c P values (2-sided) were statistically significant at <0.05
There was a statistically significant interaction (P for interaction = 0.01) by smoking for the DASH-Mellen dietary pattern score, where current smokers or those who quit <10 years ago in the highest quintile (quintile 5 vs. quintile 1) had a lower risk (HR = 0.74, 95% CI: 0.58, 0.94, P for trend = 0.002) while no association was present in never smokers or those who quit ≥10 years ago (HR = 0.97, 95% CI: 0.85, 1.12; P for trend = 0.87) (Web Table 2). There were no other significant interactions by smoking, race, body mass index, or alcohol consumption. All analyses were proportional over time (P > 0.05). Overall, in 5-year lagged analyses, the HEI-2015 and aMed remained significantly associated with PDAC but DASH-Fung and DASH-Mellen did not (Web Table 3).
DISCUSSION
In this large cohort of middle-aged and older adults, greater adherence to diet quality indices was associated with lower PDAC risk. Comparing the highest with the lowest quintiles of adherence, 4 diet quality indices—HEI-2015, aMed, and 2 versions of DASH—were associated with a 15%–18% lower risk of PDAC. For women, only aMed adherence remained significant, with a 24% lower PDAC risk.
Previous studies have examined HEI-2005 and various versions of Mediterranean diet scores and PDAC risks (25–27, 41, 42). Consistent with our findings, an earlier analysis in NIH-AARP (n = 2,383 cases) showed a 15% overall lower PDAC risk with higher adherence to HEI-2005 (26). The HEI-2015 differs from the HEI-2005 to reflect evolving dietary guidance and more specific construction of the score, including the addition of the greens and beans component (replacing dark-green and orange vegetables and legumes), total protein foods and seafood and plant protein (replacing meat and beans), fatty acid ratio (replacing oils and saturated fat), refined grains as a moderation component (replacing the adequacy component total grains), and separate components for added sugars and saturated fat (replacing the empty calories from solid fat, alcohol, and added sugars components). Our study showed the most robust association with higher adherence to the aMed score, with an 18% lower PDAC risk overall and significant inverse associations in both men and women. The aMed score uses population-specific medians, and the composition of other Mediterranean diet scores may differ across studies. An Italian hospital-based case-control study showed a statistically significant inverse association with greater adherence to the traditional (Greek) Mediterranean diet score and pancreatic cancer risk (n = 688 cases) (41). These contrast with the nonsignificant inverse associations or null associations observed in earlier analyses, including a pooled analysis of 2 Dutch cohorts examining the aMed and modified Mediterranean diet scores both with and without alcohol (n = 449 cases) (27) and an analysis in the European Prospective Investigation into Cancer and Nutrition cohort examining an adapted Mediterranean diet score without alcohol (n = 865 cases) (25). A recent analysis comparing 4 diet quality indices in the Singapore Chinese Health Study cohort and pancreatic cancer risk (n = 311 cases) suggested inverse associations with higher adherence to the AHEI-2010, aMed, and DASH-Fung, whereas the nutrient-based Healthy Diet Indicator was associated with higher risks (43). In this study of Chinese participants, pancreatic cancer incident cases included PDAC and those of unknown histology (43). Our present study included more PDAC cases than these earlier cohort studies and has more power to observe associations.
Data-driven dietary pattern approaches, including factor and principal components analyses, have shown inconsistent associations with pancreatic cancer (42, 44–46). A limitation of these approaches is that study-specific dietary patterns cannot be compared across studies. Inverse prospective PDAC associations have been observed for a priori–defined dietary pattern scores including total antioxidant capacity (47, 48), 2018 World Cancer Research Fund/American Institute for Cancer Research cancer prevention recommendations (49), and as components within healthy lifestyle scores (28, 50); the last 2 scores include both dietary and lifestyle components, such as body mass index and physical activity. Scores representing the inflammatory potential of diet have been inconsistently associated with pancreatic cancer in prospective studies (51–53).
We observed significant associations in men, but not women, when defining diet quality with the HEI-2015 and the 2 DASH scores, although the interaction by sex was not significant. This could be due to differences in self-reported dietary intake, sex-related biological effects of diet, the larger proportion of male participants in NIH-AARP, or dietary score construction. The aMed showed similar and significant associations in both men and women. The HEI-2015 and DASH-Mellen used the same cutpoints for men and women, and they had food or nutrient components that were energy-density adjusted (Table 1), while the DASH-Fung used sex-specific quintile cutpoints for all components. None of these included the alcohol component in the score. In contrast, the aMed’s simplified dichotomous scoring approach based on intake above or below sex-specific median intake in NIH-AARP and moderate intake ranges for the alcohol component might have contributed to the significant associations in both men and women. The AHEI-2010 showed no significant associations for both sexes. This may be due to construction of the dietary score, including the scoring approach (e.g., absolute intake and maximum points given for consuming no red or processed meat, sugar sweetened beverages, or fruit juices), fatty acid components (e.g., polyunsaturated fat, eicosapentaenoic acid, docosahexaenoic acid, trans-fat), or other differences compared with the HEI-2015, aMed, and DASH-Fung.
In our exploratory by-component analyses, we observed lower PDAC risk with lower consumption of red and processed meats and moderate alcohol consumption as defined by aMed but not as defined by AHEI-2010. Higher intakes of whole grains as defined by HEI-2015, AHEI-2010, and DASH-Fung—but not as defined by aMed—were associated with lower PDAC risk. The dairy and calcium components were inversely associated with PDAC as defined by HEI-2015 and DASH-Mellen, respectively; however, the low-fat dairy DASH-Fung component was not associated with PDAC. The differences in individual associations for components across the dietary scores are likely due to different cutpoints and comparison groups across the indices. For example, the aMed red and processed meat component (based on sex-specific median cutpoints) showed inverse associations. The AHEI-2010 and DASH-Fung red and processed meat component did not show associations. The AHEI-2010 scored optimal red and processed meat consumption as “none” and the DASH-Fung scored optimal red and processed meat consumption as the lowest sex-specific quintile. Compared with the other diet quality indices, the AHEI-2010 gives greater weight to fatty acids that have not been associated with PDAC in the NIH-AARP (10). Individual components of index-based diet quality are not meant to be interpreted independently, as they do not account for synergistic relationships.
Strengths of our study include its large prospective design, with dietary data collected on individuals prior to cancer diagnosis, and long follow-up time; thus, our results are less likely to be influenced by reverse causation and selection or recall biases and have internal validity. In addition to the uniform approach of calculating food components across scores following the Dietary Patterns Methods Project, dietary quality was based on public health guidelines or healthful eating recommendations that reflect a broad range of scientific evidence, including that from epidemiologic studies. Our study includes a large number of PDAC cases, as well as a wide distribution of dietary intake, providing greater power to detect differences and associations.
There are also limitations. Measurement error inherent to dietary assessment using FFQs is likely present and could result in inaccurate risk estimates. Diet was measured only at baseline; repeated measurements would increase the accuracy of the dietary assessment. As score-based dietary patterns are truncated and some individual components are dichotomized, these scores do not reflect the effects of excessive intakes of certain components (e.g., protein, total fruits including juices, or dairy products) and may not capture important information on differences in food and nutrient intake across individuals. Residual confounding and unmeasured exposures associated with both diet quality and PDAC could have influenced our observed associations. Most of the NIH-AARP Study participants are non-Hispanic White persons and our results might not be generalizable to other racial or ethnic groups. Future studies investigating dietary patterns and PDAC should include a more racially and ethnically diverse population.
In conclusion, results from this large prospective cohort support the hypothesis that greater adherence to dietary recommendations based on scientific evidence may reduce the risk of developing PDAC. Higher diet quality index scores have also been associated with lower risks of type 2 diabetes mellitus and body adiposity, known risk factors for PDAC, which could contribute to some of the associations we observe with PDAC. Diet quality represents an important potentially modifiable risk factor that could decrease the burden of pancreatic cancer.
Supplementary Material
ACKNOWLEDGMENTS
Author affiliations: Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health, Bethesda, Maryland, United States (Sachelly Julián-Serrano, Rachael Stolzenberg-Solomon); Division of Cancer Control and Population Sciences, National Cancer Institute, National Institutes of Health, Bethesda, Maryland, United States (Jill Reedy); Department of Exercise and Nutrition Sciences, Milken Institute School of Public Health, George Washington University, Washington, District of Columbia, United States (Kim Robien); and Department of Epidemiology, Milken Institute School of Public Health, George Washington University, Washington, District of Columbia, United States (Kim Robien).
This work was supported by the Intramural Research Program, Division of Cancer Epidemiology and Genetics of the US National Cancer Institute, National Institutes of Health, Department of Health and Human Services.
Data are available upon request from https://www.aarp.org/forms/research-dataset-request-form/.
The authors gratefully acknowledge Dr. Kara Michels and Dr. Barry I. Graubard (Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health) for their assistance in data analysis and interpretation. Cancer incidence data were collected by the Georgia Center for Cancer Statistics, Department of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, GA (for the Atlanta metropolitan area); the California Cancer Registry, California Department of Public Health’s Cancer Surveillance and Research Branch, Sacramento, CA; the Michigan Cancer Surveillance Program, Community Health Administration, Lansing, MI (for the Detroit metropolitan area); the Florida Cancer Data System (Miami, Florida) under contract with the Florida Department of Health, Tallahassee, FL; the Louisiana Tumor Registry, Louisiana State University Health Sciences Center School of Public Health, New Orleans, LA; the New Jersey State Cancer Registry, The Rutgers Cancer Institute of New Jersey, New Brunswick, NJ; the North Carolina Central Cancer Registry, Raleigh, NC; the Division of Health Statistics and Research, Pennsylvania Department of Health, Harrisburg, PA; the Arizona Cancer Registry, Division of Public Health Services, Arizona Department of Health Services, Phoenix, AZ; the Texas Cancer Registry, Cancer Epidemiology and Surveillance Branch, Texas Department of State Health Services, Austin, TX; and the Nevada Central Cancer Registry, Division of Public and Behavioral Health, State of Nevada Department of Health and Human Services, Carson City, NV.
Presented at the 2021 American Association for Cancer Research Annual Meeting (online), April 9–14, 2021.
The views expressed herein are solely those of the authors and do not necessarily reflect those of the Florida Cancer Data System or Florida Department of Health. The Pennsylvania Department of Health specifically disclaims responsibility for any analyses, interpretations, or conclusions.
Conflict of interest: none declared.
REFERENCES
- 1. Siegel RL, Miller KD, Fuchs HE, et al. Cancer statistics, 2021. CA Cancer J Clin. 2021;71(1):7–33. [DOI] [PubMed] [Google Scholar]
- 2. Gordon-Dseagu VL, Devesa SS, Goggins M, et al. Pancreatic cancer incidence trends: evidence from the Surveillance, Epidemiology and End Results (SEER) population-based data. Int J Epidemiol. 2018;47(2):427–439. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Ilic M, Ilic I. Epidemiology of pancreatic cancer. World J Gastroenterol. 2016;22(44):9694–9705. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Naudin S, Li K, Jaouen T, et al. Lifetime and baseline alcohol intakes and risk of pancreatic cancer in the European Prospective Investigation into Cancer and Nutrition study. Int J Cancer. 2018;143(4):801–812. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Wang YT, Gou YW, Jin WW, et al. Association between alcohol intake and the risk of pancreatic cancer: a dose-response meta-analysis of cohort studies. BMC Cancer. 2016;16:212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Jiao L, Silverman DT, Schairer C, et al. Alcohol use and risk of pancreatic cancer: the NIH-AARP Diet and Health Study. Am J Epidemiol. 2009;169(9):1043–1051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Taunk P, Hecht E, Stolzenberg-Solomon R. Are meat and heme iron intake associated with pancreatic cancer? Results from the NIH-AARP diet and health cohort. Int J Cancer. 2016;138(9):2172–2189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Rohrmann S, Linseisen J, Nöthlings U, et al. Meat and fish consumption and risk of pancreatic cancer: results from the European Prospective Investigation into Cancer and Nutrition. Int J Cancer. 2013;132(3):617–624. [DOI] [PubMed] [Google Scholar]
- 9. McCullough ML, Jacobs EJ, Shah R, et al. Meat consumption and pancreatic cancer risk among men and women in the Cancer Prevention Study-II Nutrition Cohort. Cancer Causes Control. 2018;29(1):125–133. [DOI] [PubMed] [Google Scholar]
- 10. Thiébaut AC, Jiao L, Silverman DT, et al. Dietary fatty acids and pancreatic cancer in the NIH-AARP diet and health study. J Natl Cancer Inst. 2009;101(14):1001–1011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Moeller SM, Reedy J, Millen AE, et al. Dietary patterns: challenges and opportunities in dietary patterns research an Experimental Biology workshop, April 1, 2006. J Am Diet Assoc. 2007;107(7):1233–1239. [DOI] [PubMed] [Google Scholar]
- 12. Reedy J, Lerman JL, Krebs-Smith SM, et al. Evaluation of the Healthy Eating Index-2015. J Acad Nutr Diet. 2018;118(9):1622–1633. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Krebs-Smith SM, Pannucci TE, Subar AF, et al. Update of the Healthy Eating Index: HEI-2015. J Acad Nutr Diet. 2018;118(9):1591–1602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. U.S. Department of Health and Human Services and U.S. Department of Agriculture . 2015–2020 Dietary Guidelines for Americans. 8th ed. http://health.gov/dietaryguidelines/2015/guidelines/. Accessed on February 17, 2021.
- 15. Chiuve SE, Fung TT, Rimm EB, et al. Alternative dietary indices both strongly predict risk of chronic disease. J Nutr. 2012;142(6):1009–1018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. McCullough ML, Feskanich D, Stampfer MJ, et al. Diet quality and major chronic disease risk in men and women: moving toward improved dietary guidance. Am J Clin Nutr. 2002;76(6):1261–1271. [DOI] [PubMed] [Google Scholar]
- 17. Fung TT, McCullough ML, Newby PK, et al. Diet-quality scores and plasma concentrations of markers of inflammation and endothelial dysfunction. Am J Clin Nutr. 2005;82(1):163–173. [DOI] [PubMed] [Google Scholar]
- 18. Trichopoulou A, Kouris-Blazos A, Wahlqvist ML, et al. Diet and overall survival in elderly people. BMJ. 1995;311(7018):1457–1460. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Fung TT, Chiuve SE, McCullough ML, et al. Adherence to a DASH-style diet and risk of coronary heart disease and stroke in women. Arch Intern Med. 2008;168(7):713–720. [DOI] [PubMed] [Google Scholar]
- 20. Appel LJ, Moore TJ, Obarzanek E, et al. A clinical trial of the effects of dietary patterns on blood pressure. DASH Collaborative Research Group. N Engl J Med. 1997;336(16):1117–1124. [DOI] [PubMed] [Google Scholar]
- 21. Sacks FM, Svetkey LP, Vollmer WM, et al. Effects on blood pressure of reduced dietary sodium and the Dietary Approaches to Stop Hypertension (DASH) diet. DASH-Sodium Collaborative Research Group. N Engl J Med. 2001;344(1):3–10. [DOI] [PubMed] [Google Scholar]
- 22. Vollmer WM, Sacks FM, Ard J, et al. Effects of diet and sodium intake on blood pressure: subgroup analysis of the DASH-sodium trial. Ann Intern Med. 2001;135(12):1019–1028. [DOI] [PubMed] [Google Scholar]
- 23. Morze J, Danielewicz A, Hoffmann G, et al. Diet quality as assessed by the Healthy Eating Index, Alternate Healthy Eating Index, Dietary Approaches to Stop Hypertension score, and health outcomes: a second update of a systematic review and meta-analysis of cohort studies. J Acad Nutr Diet. 2020;120(12):1998–2031.e15. [DOI] [PubMed] [Google Scholar]
- 24. Liese AD, Krebs-Smith SM, Subar AF, et al. The Dietary Patterns Methods Project: synthesis of findings across cohorts and relevance to dietary guidance. J Nutr. 2015;145(3):393–402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Molina-Montes E, Sánchez MJ, Buckland G, et al. Mediterranean diet and risk of pancreatic cancer in the European Prospective Investigation into Cancer and Nutrition cohort. Br J Cancer. 2017;116(6):811–820. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Arem H, Reedy J, Sampson J, et al. The Healthy Eating Index 2005 and risk for pancreatic cancer in the NIH-AARP study. J Natl Cancer Inst. 2013;105(17):1298–1305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Schulpen M, Peeters PH, van den Brandt PA. Mediterranean diet adherence and risk of pancreatic cancer: A pooled analysis of two Dutch cohorts. Int J Cancer. 2019;144(7):1550–1560. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Jiao L, Mitrou PN, Reedy J, et al. A combined healthy lifestyle score and risk of pancreatic cancer in a large cohort study. Arch Intern Med. 2009;169(8):764–770. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Mellen PB, Gao SK, Vitolins MZ, et al. Deteriorating dietary habits among adults with hypertension: DASH dietary accordance, NHANES 1988–1994 and 1999–2004. Arch Intern Med. 2008;168(3):308–314. [DOI] [PubMed] [Google Scholar]
- 30. Schatzkin A, Subar AF, Thompson FE, et al. Design and serendipity in establishing a large cohort with wide dietary intake distributions: the National Institutes of Health–American Association of Retired Persons Diet and Health Study. Am J Epidemiol. 2001;154(12):1119–1125. [DOI] [PubMed] [Google Scholar]
- 31. Thompson FE, Kipnis V, Midthune D, et al. Performance of a food-frequency questionnaire in the US NIH-AARP (National Institutes of Health–American Association of Retired Persons) Diet and Health Study. Public Health Nutr. 2008;11(2):183–195. [DOI] [PubMed] [Google Scholar]
- 32. Subar AF, Midthune D, Kulldorff M, et al. Evaluation of alternative approaches to assign nutrient values to food groups in food frequency questionnaires. Am J Epidemiol. 2000;152(3):279–286. [DOI] [PubMed] [Google Scholar]
- 33. Subar AF, Thompson FE, Smith AF, et al. Improving food frequency questionnaires: a qualitative approach using cognitive interviewing. J Am Diet Assoc. 1995;95(7):781–788. [DOI] [PubMed] [Google Scholar]
- 34. Reedy J, Krebs-Smith SM, Miller PE, et al. Higher diet quality is associated with decreased risk of all-cause, cardiovascular disease, and cancer mortality among older adults. J Nutr. 2014;144(6):881–889. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Michaud D, Midthune D, Hermansen S, et al. Comparison of cancer registry case ascertainment with SEER estimates and self-reporting in a subst of the NIH-AARP Diet and Health Study. J Registry Manag. 2005;32(2):70–75. [Google Scholar]
- 36. Maldonado G, Greenland S. Simulation study of confounder-selection strategies. Am J Epidemiol. 1993;138(11):923–936. [DOI] [PubMed] [Google Scholar]
- 37. Willet W. Nutritional Epidemiology. 3rd ed. New York, NY: Oxford University Press; 2013. [Google Scholar]
- 38. Hu FB. Dietary pattern analysis: a new direction in nutritional epidemiology. Curr Opin Lipidol. 2002;13(1):3–9. [DOI] [PubMed] [Google Scholar]
- 39. VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the E-Value. Ann Intern Med. 2017;167(4):268–274. [DOI] [PubMed] [Google Scholar]
- 40. VanderWeele TJ, Ding P, Marthur M. Technical considerations in the use of the E-value. J Causal Inference. 2019;7(2):1–11. [Google Scholar]
- 41. Bosetti C, Turati F, Dal Pont A, et al. The role of Mediterranean diet on the risk of pancreatic cancer. Br J Cancer. 2013;109(5):1360–1366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Bosetti C, Bravi F, Turati F, et al. Nutrient-based dietary patterns and pancreatic cancer risk. Ann Epidemiol. 2013;23(3):124–128. [DOI] [PubMed] [Google Scholar]
- 43. Luu HN, Paragomi P, Jin A, et al. Quality diet index and risk of pancreatic cancer: findings from the Singapore Chinese Health Study. Cancer Epidemiol Biomarkers Prev. 2021;30(11):2068–2078. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Chan JM, Gong Z, Holly EA, et al. Dietary patterns and risk of pancreatic cancer in a large population-based case-control study in the San Francisco Bay Area. Nutr Cancer. 2013;65(1):157–164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Michaud DS, Skinner HG, Wu K, et al. Dietary patterns and pancreatic cancer risk in men and women. J Natl Cancer Inst. 2005;97(7):518–524. [DOI] [PubMed] [Google Scholar]
- 46. Inoue-Choi M, Flood A, Robien K, et al. Nutrients, food groups, dietary patterns, and risk of pancreatic cancer in postmenopausal women. Cancer Epidemiol Biomarkers Prev. 2011;20(4):711–714. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Zhong GC, Pu JY, Wu YL, et al. Total antioxidant capacity and pancreatic cancer incidence and mortality in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial. Cancer Epidemiol Biomarkers Prev. 2020;29(5):1019–1028. [DOI] [PubMed] [Google Scholar]
- 48. Lucas AL, Bosetti C, Boffetta P, et al. Dietary total antioxidant capacity and pancreatic cancer risk: an Italian case-control study. Br J Cancer. 2016;115(1):102–107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Zhang ZQ, Li QJ, Hao FB, et al. Adherence to the 2018 World Cancer Research Fund/American Institute for Cancer Research cancer prevention recommendations and pancreatic cancer incidence and mortality: a prospective cohort study. Cancer Med. 2020;9(18):6843–6853. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50. Naudin S, Viallon V, Hashim D, et al. Healthy lifestyle and the risk of pancreatic cancer in the EPIC study. Eur J Epidemiol. 2020;35(10):975–986. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51. Zheng J, Wirth MD, Merchant AT, et al. Inflammatory potential of diet, inflammation-related lifestyle factors, and risk of pancreatic cancer: results from the NIH-AARP Diet and Health Study. Cancer Epidemiol Biomarkers Prev. 2019;28(7):1266–1270. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52. Zheng J, Merchant AT, Wirth MD, et al. Inflammatory potential of diet and risk of pancreatic cancer in the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial. Int J Cancer. 2018;142(12):2461–2470. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53. Antwi SO, Bamlet WR, Pedersen KS, et al. Pancreatic cancer risk is modulated by inflammatory potential of diet and ABO genotype: a consortia-based evaluation and replication study. Carcinogenesis. 2018;39(8):1056–1067. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.